1 Grammatical Accuracy and Learner Autonomy in Advanced ESL Writing (forthcoming in the English Language Teaching Journal, May 2006) Abstract Caroline Vickers California State University, San Bernardino Estela Ene University of Arizona This paper aims to explore advanced ESL learners' ability to make improvements in grammatical accuracy by autonomously noticing and correcting their own grammatical errors. In the recent literature in SLA, it is suggested that classroom tasks can be used to foster autonomous language learning habits (c.f., Dam 2001). Therefore, it is important to consider classroom tasks that encourage autonomous language learning behavior. Working with 13 advanced ESL composition students, we engaged the subjects in an explicit task in which they compared their own use of grammatical form in their own written output to the use of grammatical form as used in a text written by a native speaker. Based on the comparison between their own written output and the native speaker text, subjects subsequently corrected their grammatical errors. Results suggest that such a comparison task is beneficial in allowing learners to make gains in grammatical accuracy. Introduction The issue of learner autonomy is of general concern in the Second Language Acquisition (SLA) community. The International Association of Applied Linguistics (AILA) Review has recently dedicated a special issue to learner autonomy in second language (L2) learning, demonstrating the validity and benefit of autonomous learning environments (Dam 2001).
2 Importantly, learner autonomy is conceptualized not as simply setting learners to tasks such as interactive computer tasks, or as declaring the instructor obsolete, but as a pedagogical ideology in favor of turning control of the task of learning over to the learner not only to promote good language learning habits but as a way of empowering the learner (Benson and Voller 1997). The benefits of learner autonomy have been recognized in the field of L2 writing as well. Based on her experience as a teacher and researcher in the area of L2 writing, Ferris recommends to L2 writing learners "be aware of your own individual error patterns" (2002: 87). Generally, teachers and researchers recognize part of the role of instruction is to promote beneficial language learning habits. For instance, Lightbown and Spada (1999) indicate that one important aspect of classroom language teaching is helping learners to notice form in the language, which not only allows them to become more accurate regarding the form in focus but also promotes a language learning skill that learners can carry with them beyond the present instructional situation. It seems that encouraging learner autonomy is increasingly recognized as a beneficial practice to promote good language learning. Therefore, it is necessary to explore language learning tasks that encourage learner autonomy and that also lead to gains in accuracy in the second language. The present study is an attempt to find such a task suitable for the advanced ESL writing classroom. Learner Autonomy and Second Language Acquisition Research in second language acquisition (SLA) emphasizes the need to help L2 learners notice their own L2 use in comparison to the use of the target language as produced by native speakers. Doughty and Williams state that "one of the central issues in Focus on Form (FonF) research is how to lead the learner's attention to a linguistic mismatch between IL (interlanguage)
3 and TL (target language)" (1998: 238). It is argued that the recognition of this mismatch between the learner's IL and the TL promotes language learning. The FonF literature seems to imply that the recognition of the mismatch is an autonomous process for language learners. Specific tasks have been identified in the SLA literature that demonstrate the gains that learners make when they autonomously locate a mismatch between their IL and the TL, most notably when engaged in tasks that involve the written modality. Reformulation as described by Cohen (1989) has proven productive as a way of providing correction on L2 writing. In a recent study on reformulation, Qi and Lapkin (2001) note the ability of ESL writers to notice grammatical errors in their own written output when comparing those errors to reformulations of their written texts by native speakers. In Qi and Lapkin's study, the subjects expressed awareness of a difference between their own use of the language and the use of the language by native speakers in the reformulations. As Qi and Lapkin argue, "this demonstrates that the noticed features of the modeled TL behavior were being constantly compared to the learner's own written text and that the subject's own recent experience of output was an important factor in influencing what he/she noticed" (2001: 290). Having learners compare their texts to native speaker reformulations seems to allow learners to be autonomous in their ability to find their own output errors. However, the text reformulation requires assistance from a native speaker, which may not be available or practical in all language learning contexts. It is possible for learners to notice the mismatch by comparing their own output to reading passages rather than native speaker reformulations. Izumi (2002) demonstrates that when language learners have a chance to compare their own written output to a reading passage, they benefit in terms of their ability to use the form grammatically. In Izumi's study, the use of the
4 relative clause was typographically enhanced within the reading passage by using different font types to make the relative clause more noticeable in the context of the reading passage. Learners specifically benefited from producing written output and later being exposed to the typographically enhanced reading passage in terms of their ability to subsequently produce relative clauses accurately. The acquisition of the relative clause in Izumi's study was, therefore, an autonomous process for the learners in which they compared their own language use to a typographically enhanced text. The implicit task described by Izumi is effective in terms of promoting noticing and acquisition. Doughty and Williams describe implicit tasks, such as the typographically enhanced reading passage described in Izumi's study, as having "the aim to attract learner attention to avoid metalinguistic discussion" and explicit tasks, such as rule explanation followed by practice, as having "the aim to direct learner attention and to exploit pedagogical grammar in this regard" (1998: 232). Importantly, Doughty and Williams find a place for both implicit and explicit tasks in the language classroom as well as techniques that are somewhere in between on a continuum between implicit and explicit. It seems that learners can make grammatical gains autonomously by engaging in implicit tasks as demonstrated by Izumi. However, as noted by Leki (1991), college level second language writers prefer explicit error correction from their instructors. Such correction contributes to motivation and confidence in the instructor (Ferris 2002:8). The question arises whether an explicit error correction task can be accomplished autonomously and allow learners to make gains in accuracy. The purpose of this study is to investigate whether learners make grammatical gains when engaging in an autonomous self-correction task that directs learner attention to form explicitly. Specifically, we examine whether classroom language learners can autonomously
5 notice and autonomously correct their own grammatical errors by 1) reading an explanation of the target form to direct learner attention to the form and 2) by comparing the use of a target form in their own writing to a text containing the typographically enhanced target form. We are interested in whether such an explicit self-correction task leads to improvements in grammatical accuracy. Finding the Mismatch in the Advanced ESL Composition Classroom Subjects The subjects for this study (n=13) were advanced learners of English enrolled in a university composition program in the southwestern United States. They were enrolled in two ESL composition courses which are the equivalent of freshman composition for native speakers of English. The students in the composition program are generally enrolled as undergraduate students in the university, and take freshman composition during their first year in residence. The subjects came from a variety of native language backgrounds, including Japanese, Spanish, Arabic, Hindi, Portuguese, Mandarin, Indonesian, Greek, Tamil, and Urdu. The ages of the subjects ranged from 17 to 35. All of the subjects had achieved at least a 500 on the TOEFL and had had at least four years of English language instruction at the time of the study. The two courses in which this research was conducted were the same level, and two separate instructors taught the two courses in which the data was collected. Those learners who scored above 90% on the Pretest did not participate in this study as they were deemed to have mastered the target form. No learner scored below 45% on the Pretest. Therefore, all learners who participated in the study were considered developmentally ready to learn the form. There
6 were a total of 13 learners out of 20 who scored between 90% and 45%. Those 13 learners were selected as subjects for the study. Method We investigated the ability of the subjects to locate the mismatch between the use of the past hypothetical conditional (e.g., If I had gone to the baseball game, I would have seen Hank Aaron hit a homerun.) in their own written output and the use of the form in a text written by one of the researchers, a native speaker. We chose to study the learners' use of the past hypothetical conditional because it is an advanced form that either did not appear in the learners' writing or appeared with errors. Furthermore, the past hypothetical conditional is commonly used in American academic discourse, especially in persuasive writing. It is, therefore, an important form to acquire as learners study the conventions of American academic discourse. The subjects in this study were enrolled as undergraduates in an American university and would be expected to use the conventions of American academic discourse in their tenure as American university students. The Pretest asked learners to produce sentences containing the past hypothetical conditional in response to 10 picture based production items such as the item in Figure 1. Based on the Pretest, we determined that learners made several different types of errors using the past hypothetical conditional. For instance, some of the errors were on the if-clause (If I had gone to the game). These errors involved the use of another tense besides the past perfect. Other errors were on the result-clause (I would have seen Hank Aaron hit a home run). These errors were of various types such as would see, saw, and would seen. Based on the Pretest, the average score of correct usage in both the if-clause and the result-clause was 68%.
7 (Insert Figure 1 here) After determining that learners made errors using the past hypothetical conditional, we assigned the learners a series of tasks to encourage them to locate a mismatch between their own IL use of the past hypothetical conditional and TL use of the form as it appeared in a 400 word reading passage written by the native English speaking researcher for the purpose of this study. Each of the following tasks were completed on a different day within a 10 day period. Day 1. Learners wrote a paragraph based on a prompt. The prompt described a situation in which their school hired a terrible teacher the month before. It then directed learners that they should hypothesize about serving on the committee that hired the terrible teacher. At the end of the prompt, the learners were provided with phrases such as if I, if the teacher, if the committee that they were asked to include in their paragraph. Day 2. Learners read a rule explanation of the past hypothetical conditional followed by a 400 word text in which various uses of the past hypothetical conditional had been bolded to make the form salient in the text. The researchers then returned the paragraphs that the learners had written on Day 1. At the bottom of the paragraphs that the learners had produced on Day 1, the researchers had written the number of errors the learners had made using the past hypothetical conditional. The learners were then encouraged to compare their own use of the past hypothetical conditional in the paragraph that they wrote on Day 1 to its use in the bolded text in order to correct their errors.
8 Day 3. Learners wrote a paragraph based on a different prompt. The prompt described a situation in which a businessperson named George missed an important meeting in Los Angeles because his car broke down on the 10 hour trip there the week before. It then directed learners that they should hypothesize about planning the trip for George the week before. As in Day 1, the learners were prompted with phrases such as if I, if George, if the car. Results One week after learners had completed all of the tasks, they were again tested on their use of the past hypothetical conditional in Posttest 1. The tests consisted of 10 picture based production items as shown in Figure 1. This time the average score was 93.05%, much higher than on Pretest given before they completed the assigned tasks. The learners were also tested 5 weeks later on Posttest 2 and obtained an average score of 92.46%. The Pretest, Posttest 1, and Posttest 2 were the same picture based production format, but the content of each item on the test was different. The learners performed better on the tests after they completed the assigned series of tasks than before, even 5 weeks after they completed the tasks. Therefore, the grammatical gains were both immediate and maintained. Examining the learners' corrected paragraphs, it is clear that the learners were able to self correct their use of the past hypothetical conditional by finding the mismatch between their use of the form on the one hand, and the use of the form in the typographically enhanced reading passage on the other hand. The learners crossed through the error and wrote above it the correct
9 form. It seems that engaging in this explicit self-correction task allowed learners to use the past hypothetical conditional more accurately. Considerations for Classroom Practice Explicit self-correction seems to be effective in terms of gains in grammatical accuracy. Dam (1995) stresses the importance of raising learner awareness of the learning process and of the system of the target language. Although the present study does provide evidence that engaging learners in an autonomous correction task allows for greater grammatical accuracy, it does not provide evidence that such an autonomous task raises learner awareness of the learning process. Further research should address systematically whether engaging learners in autonomous tasks is beneficial in terms of greater awareness of the learning process. However, the finding in this study that explicit self-correction allows for greater grammatical accuracy has important classroom implications. When devising autonomous language learning tasks, it is important to consider the materials used within the task as well as the population of learners engaging in the task. The first important consideration is the text used for comparison purposes. In the present study, we used a relatively simple text for advanced ESL learners. Ferris (2002) suggests using authentic texts which contain familiar content or content in which the grammar and vocabulary are relatively simple so that learners can scan the texts to notice a particular form. It is important not to use texts that are difficult and unfamiliar to the learners, as their attentional resources would be allocated to the content of the text rather than to the use of grammatical form in the text. Harley (1994) demonstrates that it is important that content not interfere with the ability to
10 focus on form. Therefore, texts used for the purpose of a focus on grammatical form should be appropriate to the task. It is interesting to consider other ways to draw learners' attention to grammatical form within texts that might also be effective in the classroom. Some suggestions are as follows: 1) As in our study, the teacher highlights the use of the target grammatical form in the native speaker text. 2) Learners underline the form in the native speaker text rather than the teacher highlighting the form. 3) Learners skim the native speaker text with an eye to locating the use of the grammatical form in the text. The above methods might allow learners to locate the grammatical form for the purpose of finding the mismatch between the use of the grammatical form in the native speaker texts and their own use of the grammatical form in their own written output. Classroom research could be beneficial in devising various effective classroom tasks that allow learners to locate and correct their own grammatical errors by comparing their output to native speaker input. It is important to note that the learners in the present study were from a unique population of second language learners. These language learners were advanced but also experienced. The fact that they were experienced language learners might have meant that they were prone to notice grammatical form in the text. Prompting by the researchers to notice the past hypothetical conditional might have set them into motion engaging in a task that they were accustomed to performing, though they may not have necessarily noticed the past hypothetical conditional without some encouragement. Further, many of the subjects probably had exposure to the target
11 form in previous instruction, and they might have run across the form during the time of the study in their daily lives as university students in the United States, as can be expected in a context of advanced ESL learners. Therefore, we cannot claim that the method described in this study was the only factor in the subjects' improvement. We can say, however, that these subjects had no other instruction on the use of the target form during the time of the study and that these subjects improved their use of the target form from Pretest to Posttest 1 and that they were able to maintain those gains at Posttest 2. The motivation level of these learners has to be considered as well. As learners in advanced ESL composition, they were beginning their undergraduate studies in an American university. These learners were, therefore, invested in their language learning because they had to perform on par with native English speakers as undergraduates in the university. Though native speakers were not in the ESL composition courses with these learners, they were in all of their other courses. Further, in the other courses, non-native errors may not be tolerated. These learners were aware of their status and were therefore motivated to improve their ability to communicate in English. Experienced, motivated language learners may be better able to autonomously correct their own errors than less experienced and less motivated language learners. Each instructional context should be assessed in terms of the learners' ability and motivation to participate in autonomous language learning tasks such as that presented in this paper. Conclusions The subjects in this study made gains in terms of accurate use of the past hypothetical conditional by comparing their own use of the past hypothetical conditional to its use in an
12 authentic text, and they sustained these gains. Therefore, in the advanced ESL composition classroom context, explicit self-correction seems to be effective. The learning benefit resulting from engaging in explicit self-correction suggests that learner autonomy is viable at least for advanced ESL learners. Classroom research should further explore learner autonomy in other language learning contexts. The teacher's role in helping learners to notice the mismatch most likely depends on the learners' experience with language learning as well as their motivation as language learners. In some contexts teachers may need to be more involved with learners' efforts to notice grammatical form, while in other contexts the learners may benefit more robustly from autonomous tasks. These decisions ultimately must be made by individual teachers as they understand the needs of their learners most fully.
13 References Benson, P. and P. Voller (eds.). 1997. Autonomy and Independence in Language Learning. New York: Longman. Cohen, A. 1989. 'Reformulation: A technique for providing advanced feedback in writing'. Guidelines: A Periodical for Classroom Language Teachers 11/2: 1-9. Dam, L. 1995. Learner Autonomy 3: From Theory to Classroom Practice. Dublin: Authentik. Dam, L. 2001. Learner Autonomy: New Insights. AILA Review 15. Doughty, C. and J. Williams. 1998. 'Pedagogical choices in focus on form' in C. Doughty and J. Williams (eds.): 1-12. Doughty, C. and J. Williams (eds.). 1998. Focus on Form in Classroom Second Language Acquisition. New York: Cambridge University Press. Ferris, D. 2002. Treatment of Error in Second Language Learner Writing. Ann Arbor: The University of Michigan Press. Harley, B. 1994. 'Appealing to consciousness in the L2 classroom'. In J. Hulstijn & R. Schmidt (eds.): 1994. Consciousness in Second Language Learning. AILA Review 11: 578-68. Izumi, S. 2002. 'Output, input enhancement, and the output hypothesis: An experimental study on ESL relativization'. Studies in Second Language Acquisition 24/4: 541-577. Leki, I. 1991. 'The preference of ESL students for error correction in college-level writing classes'. Foreign Language Annals 24: 203-218. Qi, D.S. and S. Lapkin. 2001. 'Exploring the role of noticing in a three-stage second language task'. Journal of Second Language Writing 10/4: 277-303.
14 Figure 1. Sample Test Item: Pretest, Posttest 1 and Posttest 2. Jane did not go out to lunch with her friends last week because she did not have any money. (last week) make money go out to lunch If